Well, regarding the study in the OP, I've been losing some weight recently, even while still eating McDonalds and such. A triathlete friend of mine told me about an application called myfitnesspal, and I started using it. I'd always heard tracking your food and calories was helpful, but before these new apps getting the calorie count for foods was a major PITA. But using this app I've been able to drop 14 lbs. Still got about 30 to go. But I eat McDonald's, pizza, Taco Hell, etc. as well as healthier prepared at home foods and restaurants like Panera and Carrabbas. Basically as long as the calorie count stays below my target for losing a lb. every couple of weeks, I'm good. Doesn't matter what the food is. And as ElizabethT says, if I do some exercise or physical labor I get to eat some additional calories.
I've stopped logging my food recently what with our southern migration, but I plan to start back up once we're settled. I could decrease the calories and lose weight faster, but I figure it took me almost 40 years to get here, so taking a year or so to get back is reasonable. It lets me still enjoy my life food-wise, while making good incremental progress.
I don't know if the study I posted is legit or not, as far as who eats the higher percentage of junk food. But I do know that the only way to lose weight is to take in fewer calories than the number that would maintain or increase your weight. I really don't think it matters where the calories come from, as long as you avoid scurvy and rickets and such. Personally I try to stay away from too much bread and starch, but only because they stimulate me to eat a ton more of them. Even as a diabetic I don't see any particular impact on my blood sugar when I eat bread and starches, as long as I stay below my calorie target. That's been my experience, anyway.